Postavel

ReviewAudited by ClawScan on May 10, 2026.

Overview

Postavel appears to be a real social-media MCP integration, but review is advised because its setup can run unpinned remote installer code with elevated privileges and the connected agent can publish or manage public social posts.

Install only if you trust Postavel and the mcporter installation source. Prefer a reviewed, pinned mcporter install over the curl-to-bash or sudo latest-binary path. After connecting, use a least-privileged Postavel account and require confirmation before publishing, deleting, auto-approving, or bulk-approving social media posts.

Findings (5)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

Running the recommended setup command could execute unreviewed remote code on the user's machine.

Why it was flagged

The skill recommends executing a remote installer directly in a shell. The remote script is not pinned or checksum-verified in the artifacts, so the user would run code whose contents may differ from the reviewed files.

Skill content
curl -fsSL https://postavel.com/install-mcp | bash
Recommendation

Avoid curl-to-bash installation where possible. Install mcporter from a trusted package manager or review and pin the installer contents before running it.

What this means

A compromised download source or changed latest release could install a malicious executable with elevated local-system impact.

Why it was flagged

The installer can download the latest mcporter binary and place it in a system executable path with sudo, without a pinned version, checksum, or signature verification.

Skill content
sudo curl -L -o /usr/local/bin/mcporter "$URL"
sudo chmod +x /usr/local/bin/mcporter
Recommendation

Pin mcporter to a specific verified release and verify checksums or signatures before installing, especially when writing to /usr/local/bin with sudo.

What this means

If the agent misunderstands a request, it could create, publish, approve, or delete social content in the user's connected Postavel workspace.

Why it was flagged

The MCP tools can publish, approve, and delete posts. This is expected for a social media management skill, but these are high-impact actions affecting public or business accounts.

Skill content
create_post ... status ... 'published' ... auto_approve ... Immediately approve ... approve_post ... delete_post
Recommendation

Require explicit user confirmation before publishing, auto-approving, bulk-approving, or deleting posts.

What this means

The agent can act within the user's Postavel permissions, including workspace and brand access that may reach connected Facebook, Instagram, or LinkedIn accounts.

Why it was flagged

The setup flow grants the AI assistant OAuth-backed access to the user's Postavel account. This is purpose-aligned, but it is still delegated account authority.

Skill content
Authorize the AI assistant to access your account
Recommendation

Use the least-privileged Postavel role that supports your needs and revoke OAuth access from Postavel settings when no longer needed.

What this means

Content and account metadata used in prompts may be transmitted to the Postavel MCP service to perform requested actions.

Why it was flagged

The skill routes agent actions through an external MCP server. This is the core integration path, but users should understand that post content, scheduling details, workspace identifiers, and media URLs may be sent to Postavel.

Skill content
MCP Server URL: `https://postavel.com/mcp/postavel`
Recommendation

Only send content and media URLs you are comfortable sharing with Postavel and verify the MCP server URL before authenticating.